Downloads · 30 days
27
4% of all-time downloads
YoozLabs/Qwen3.5-0.8B-qat-GGUF
Qwen3.5-0.8B-qat-GGUF is a text generation model from YoozLabs. Use it when you need the model to write or continue text. The card lists the license as apache-2.0.
Context limit: 8K. This build is validated healthy up to 8,192 tokens. Beyond that (tested to 32K on the MLX build of the same weights) it regresses on multi-hop retrieval -- NIAH single-needle drops to 0.03 accuracy,…
Downloads · 30 days
27
4% of all-time downloads
All-time downloads
683
Public
Repo size
501 MB
Likes
1
Public
Click a slice to open those files.
.gguf501 MB · 100%
From the Hugging Face model README
Context limit: 8K. This build is validated healthy up to 8,192 tokens. Beyond that (tested to 32K on the MLX build of the same weights) it regresses on multi-hop retrieval -- NIAH single-needle drops to 0.03 accuracy, below the untrained 4-bit baseline. Do not use this build past 8K context. If you need long context at this size class, use Qwen3.5-4B-qat-GGUF instead, which is clean through 32K. Full breakdown in Validation below.
The GGUF (Q4_0) release of Yooz Labs' quantization-aware-trained (QAT) Qwen3.5-0.8B -- the same weights as Qwen3.5-0.8B-qat-lean-4bit-mlx, converted for llama.cpp. Stock 4-bit PTQ builds of this model damage its aggregation abilities (variable-tracking collapses from a healthy score to near-zero on the MLX product grid); this build is trained to tolerate 4-bit quantization via knowledge distillation from the full-precision bf16 model, so the quantized weights reproduce the original model's behavior instead of trading one task for another.
Qwen/Qwen3.5-0.8Bllama.cpp convert_hf_to_gguf.py + llama-quantize
to Q4_0 (build e3546c7). The converter's block count is set to 24
layers, matching this model's actual depth.benchmarks/ (infinite repo, Yooz Labs).The comprehensive A/B (needle-in-a-haystack, RULER-style aggregation, MRCR) was run on the MLX build of these exact weights, since that is where the full n=30, two-context-length harness lives; see Qwen3.5-0.8B-qat-lean-4bit-mlx for the full table and chart. Summary:
| Task | Context | 4-bit PTQ, same grid | Yooz QAT 4-bit |
|---|---|---|---|
| NIAH single | 8K | 1.00 | 1.00 |
| NIAH multikey | 8K | 1.00 | 1.00 |
| Variable tracking | 8K | 0.23 | 0.78 |
| Frequent words | 8K | 0.28 | 0.36 |
| MRCR 2-needle | 8K | 0.80 | 0.67 |
| NIAH single | 32K | 1.00 | 0.03 |
| NIAH multikey | 32K | 1.00 | 0.50 |
| Variable tracking | 32K | 0.49 | 0.16 |
| Frequent words | 32K | 0.33 | 0.31 |
| MRCR 2-needle | 32K | 0.67 | 0.64 |
This GGUF file was separately spot-checked through the actual llama.cpp Q4_0 path (not just the MLX grid) at 8K context, n=30, on an RTX 4090 (13th Gen Intel Core i9-13900K, Ubuntu 24.04, llama.cpp build e3546c7):
| Task | Context | Yooz QAT 4-bit (this GGUF file) |
|---|---|---|
| Variable tracking | 8K | 0.88 |
| MRCR 2-needle | 8K | 0.69 |
Consistent with the MLX numbers above -- the fix survives re-quantization to a different format. As with the MLX build, do not use this file past 8K: the 32K regression was characterized on the MLX build of the same weights and there is no reason to expect Q4_0 fixes it.
llama-cli -m Qwen3.5-0.8B-qat-Q4_0.gguf -p "List the first five prime numbers." -n 64
Or with llama-server for an OpenAI-compatible endpoint:
llama-server -m Qwen3.5-0.8B-qat-Q4_0.gguf --port 8080
An MLX build of the same QAT weights is available at YoozLabs/Qwen3.5-0.8B-qat-lean-4bit-mlx.
Maintained by Yooz Labs -- sovereign, on-device AI for macOS. Everything we ship runs locally; nothing goes to a cloud. This build comes out of our long-context research line, where we found and quantified the PTQ aggregation crater -- and then trained it away.
Questions, issues, or corrections: [email protected]
Apache 2.0, inherited from the base checkpoint Qwen/Qwen3.5-0.8B. Qwen is
a model family by Alibaba Cloud; this is an independent, unofficial QAT
build, not distributed, endorsed, or certified by Alibaba.